earth2studio-discover

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). SKILL.md requires fetching live documentation pages from public URLs (e.g., nvidia.github.io earth2studio modules/examples) at runtime, so the agent ingests outsider-authored free text from web pages into the LLM context.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly requires fetching live documentation and lexicon source files at runtime (e.g., https://github.com/NVIDIA/earth2studio/tree/main/earth2studio/lexicon and multiple https://nvidia.github.io/earth2studio/... pages) and uses those fetched contents to determine and verify recommendations, so these external URLs directly control the agent's prompts/decisions.